Research map / Image generation / Image restoration
CRD-Net: Frequency-Adaptive Feature Injection and Change Decoupling for Building Damage Assessment
Yao Zheng, Yuanxin Ye, Tan Shu, Liwei Cai
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| fine-tune DINOv3 with LoRA instead of keeping it frozen (ablation) | CRD-Net baseline · xBD | Joint F1: 77.56 -> 79.55 | Improved |
| use LoRA instead of Adapter for DINOv3 fine-tuning (ablation) | DINOv3 · xBD | Joint F1: 78.17 -> 79.55 | Improved |
| add SFAFE frequency-adaptive feature enhancement (ablation) | DINOv3-LoRA · xBD | Joint F1: 79.55 -> 80.83 | Improved |
| add CRD2 decoder (OSD + PHCL) (ablation) | DINOv3-LoRA · xBD | Joint F1: 79.55 -> 80.70 | Improved |
| add OSD only, without PHCL (ablation) | DINOv3-LoRA · xBD | Joint F1: 79.55 -> 80.16 | Improved |
| add PHCL only, without OSD (ablation) | DINOv3-LoRA · xBD | Joint F1: 79.55 -> 80.50 | Improved |
| add CRD2 on top of LoRA baseline (ablation) | DINOv3-LoRA · xBD | building localization F1: 86.74 -> 86.22 | Hurt |
| combine SFAFE with CRD2 (full model) versus SFAFE only (ablation) | CRD-Net · xBD | Joint F1: 80.83 -> 81.26 | Improved |
| CRD-Net versus best prior method FlowMamba | CRD-Net · EBD destroyed class | F1: DAHiTra 64.50 vs CRD-Net 61.58 | Mixed |
| CRD-Net versus FlowMamba on xBD localization | CRD-Net · xBD | building localization F1: 87.66 -> 87.10 | Hurt |
Closest papers by meaning
- Generalizable Disaster Damage Assessment via Change Detection with Vision Foundation Model (AAAI 2025 · similarity 0.84)
- DisasterM3: A Remote Sensing Vision-Language Dataset for Disaster Damage Assessment and Response (NeurIPS 2025 · similarity 0.81)
- 3D Building Reconstruction from Monocular Remote Sensing Images with Multi-level Supervisions (CVPR 2024 · similarity 0.81)
- Feature Spectrum Learning for Remote Sensing Change Detection (CVPR 2025 · similarity 0.80)
- DISCount: Counting in Large Image Collections with Detector-Based Importance Sampling (AAAI 2024 · similarity 0.80)
- Continuous Adverse Weather Removal via Degradation-Aware Distillation (CVPR 2025 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
